The figure above represents the 2007 mars rover prototype as it existed in late 2000. The suspension from each side connects to the chassis at one point, and the two sides are differentially linked. This means that when you rotate one arm one way, the other rotates the other way. So, as the rover negotiates large obstacles, the chassis stays relatively level and stable. This is called the "Rocker-Bogie" suspension system & was invented by Don Bickler at JPL. It was used on the Sojourner rover in '97, and will be used on the dual rovers to be launched in 2003, the Mega Rover in 2007, and the subsequent Mars Sample Return mission. It is the best combination of simplicity, robustness, and high mobility that has been designed to date. It is limited to low-speed operation at this point. Traveling at high speeds on Mars (unmanned) just isn't feasible because the necessary information can't be transmitted and processed fast enough.

There are a number of ways to steer this vehicle. The most obvious is skid steering (tank style). This works fine, but is lacking from a navigation and control standpoint. The rate of turn for a certain number of wheel revolutions is highly dependent on the characteristics of the surface you're on. That is, you can't command the rover to turn the wheels for x number of seconds or revolutions and know where you're going to end up with enough precision.

The accepted best way of performing a traverse (going from point A to point B) entails traveling in straight lines connected with turns in place to correct heading. The rovers do not travel and turn at the same time. This could be done, but would require an exponential increase in the complexity of the control algorithms and wouldn't really be worth the effort and risk.

The method of steering that seems to be the best compromise is an Ackerman-type, 4-wheel-steered turn in place. Ackerman steering geometry means that the center of the turn radius of each wheel is in the same place. The front wheels of your car work this way. The wheels on the inside of the turn travel a shorter distance than those on the outside (tighter turn radius) so are placed at a tighter angle to prevent skidding. For the rover, the four corner wheels turn. The wheels at the front corners toe in and those at the rear corners rotate the opposite way, such that the center of the turn radius of each wheel is at the same place as the others. The center wheels do not turn, but due to their location they do not appreciably detract from steering performance.

The question here lies in the method of actually turning the wheels. Mass is always a huge issue in the design of flight hardware. In addition, mass and its location has a large effect on the mobility of the vehicle. Any mass devoted to steering hardware is located at each wheel, in the critical "unsuspended" areas. The drive and steering actuators (motor/gearbox combinations) are located within or near each wheel. The actuators are heavy units relative to most other subsystems. A full third of the Sojourner rover's mass was in its wheels and drive and steering actuators. Obviously, anything that can be done here for mass savings would be a significant improvement.

The idea is to try to design the steering geometry to eliminate the steering actuators altogether. This is not a developed idea at all, and that's where we come in. The rover prototype has an early, primitive attempt at this. If you look at the rover prototype picture, you can see a number of notches arranged radially around the outside (see figure below) of the wheel areas. These are locking locations for the various wheel angles.

The steering kingpin (axis) is located behind the wheel a distance. A pin is manually removed from a notch, and the drive actuator runs the wheel around this axis until it is in the right location, at which point the pin is replaced. With the wheels re-locked into place, the drive actuators are run again to actually turn the rover, and then the process is reversed to straighten the wheels again. As you can tell, this is a cumbersome process, but it eliminates the need for separate steering actuators. Needless to say, there is much room for improvement, and it is potentially very important. In automotive steering systems, very small changes in the kingpin angle and location, wheel angle, etc. translate into large changes in performance. We will study the effects of these changes on our type of steering, through computer modeling, testing physical models, etc. We also need to make the system much more compact, and eliminate the need to manually manipulate the locking pins.